I think the entire core of his argument is a sleight-of-hand between "improbable" and "the kind of absurd improbability involved in Pascal's wager", without even (as others have pointed out) giving any arguments for why it's improbable in the first place.